Carl-Ivar Brändén 1934–2004
Total Page:16
File Type:pdf, Size:1020Kb
OBITUARY Carl-Ivar Brändén 1934–2004 Shuguang Zhang, Alexander Rich, Joel L Sussman & Alan R Fersht Carl-Ivar Brändén of the Karolinska Institute died on April 28, 2004, higher language so the whole program was written in machine code. two weeks short of his 70th birthday. Carl (“Calle” to his friends and It took six months for Brändén to write a detailed flow chart, four family) was born in a tiny village in Lappland in the far north of months for Åsbrink to write the machine code and one year for them Sweden. His father was the local schoolteacher and Carl spent his first to debug the program. Over the next ten years this program was used six years at school under his own father’s supervision. There were only by the entire Scandinavian crystallographic community and was top- 15 children in the school, all in the same classroom, so when one age rated in the use of computer time for both BESK and its much group was in session, the other pupils were studying on their own. He improved successor FACIT. learned at an early age to concentrate on his work and ignore the noise In order to graduate, Carl had to take another course. His choice was around him. The village was poor and scholarly pursuit was unheard biochemistry. This decision completely changed his plans for the of. The climate consisted of nine months of winter and three months future, because he realized that he could apply his knowledge of crys- of cold wind; but nature was wonderful with beautiful lakes filled with tallography to scientifically important and intellectually stimulating lots of fish and deep forests full of berries and mushrooms and trees to problems in biology. He wanted immediately to move into the field of climb. Carl had the marvelous experience of being surrounded by a protein crystallography. The obvious place to learn was the MRC http://www.nature.com/natstructmolbiol herd of 10,000 reindeer. Laboratory for Molecular Biology (LMB) in Cambridge, UK. As Carl left home at the age of 13 with the intent of becoming a school Brändén himself put it, one year as a postdoctoral fellow in the labora- teacher. He won a scholarship to study mathematics and chemistry at tory of Sir John Kendrew at LMB converted Carl from a rather igno- Uppsala University. While studying in the chemistry laboratory, he was rant chemist to a devoted molecular biologist and he studied the invited by the acting chairman of the inorganic chemistry department, structure and function of proteins from that point on. Ingvar Lindqvist, to do a Ph.D. in his group. At the time, it was unusual LMB was in those days one of the most exciting and stimulating for a major in mathematics to work in chemistry, but since Lindqvist’s places for a young postdoctoral fellow. Max Perutz had built up the research focused on X-ray crystallography of small inorganic com- laboratory with an incredible ability to choose the right people and let plexes, he was eager to obtain students with a mathematical back- them work independently. Perutz himself was working on the high- ground. Carl was extremely flattered and accepted without hesitation. resolution structure of hemoglobin. Jacques Monod visited to lecture As Carl aptly put it, he never had a single boring day thereafter. and discuss with Perutz his latest development on the theory of In 1960, Carl attended a summer school in Manchester on modern allosteric transitions, for which the conformational changes in hemo- methods of X-ray crystallography organized by D.W.J. Cruickshank. At globin played an important role. Francis Crick and Sydney Brenner © 2004 Nature Publishing Group the time, determining the structure of even a simple small molecule were working on the genetic code and attracted an endless stream of required intense computing power. Structure factors were calculated stimulating visitors. Fred Sanger’s group was sequencing several large using mechanical calculators to add precomputed sine and cosine func- proteins and interacted frequently with the structural biologists. tions listed on so-called Beevers-Lipson strips. Most people then used a Aaron Klug had just moved in from London and was developing his homemade analog computer for Fourier summations, the Hägg- methods of optical filtering and three-dimensional reconstruction Laurent machine, where three-dimensional Fourier sums were broken from electron micrographs. The daily interactions with the postdocs down into three successive one-dimensional summations. In a series of in structural biology provided Carl with a network of lectures Cruickshank described how the method of least squares could colleagues who later became lifelong friends—Michael Rossmann, be applied to the refinement of X-ray structures and also outlined his David Blow, Lubert Stryer, Richard Dickerson, Allen Edmundson, Ken program for one of the first digital computers in the United Kingdom. Holmes, and Herman Watson, among others. Inspired by these lectures, Carl developed a collaboration with Stig At the LMB, Carl rapidly engaged himself in writing a computer Åsbrink of Stockholm University to write a similar program for the program for the refinement of myoglobin by the least squares first Swedish electronic computer, BESK. This computer had a mem- method. The structure had been solved a few years earlier and higher- ory of 1,000 words, equipped with a larger magnetic drum memory, resolution data had been collected by David Phillips at the Royal and used paper tape for input and output. There was no compiler for a Institution in London using a newly developed linear diffractometer. John Kendrew and Herman Watson were using these data in combi- S.Z. is at the Center for Biomedical Engineering and the Center for Bits nation with approximate coordinates of the myoglobin structure for and Atoms, Massachusetts Institute of Technology, Cambridge, refinement by Fourier synthesis, but better methods were needed. Massachusetts 02139-4307, USA (e-mail: [email protected]), A.R. is at Brändén and Holmes together designed a program using exact con- the Department of Biology, Massachusetts Institute of Technology, straints and a 6 × 6 block diagonal matrix. It turned out to be a non- Cambridge, Massachusetts 02139-4307, USA, J.L.S. is at the Department trivial problem to adopt this method to available computers. Even of Structural Biology, Weizmann Institute of Science, Rehovot 76100, though they were using the fastest available machines in the United Israel (e-mail: [email protected]), and A.R.F. is at Cambridge Kingdom, IBM 709 and later 7090, they had to invent several tricks to University Chemical Laboratory, Lensfield Road, Cambridge CB2 1EW, reduce computing time and storage space. The final version of the UK (e-mail: [email protected]). program required 12 magnetic tape stations for intermediate storage. 490 VOLUME 11 NUMBER 6 JUNE 2004 NATURE STRUCTURAL & MOLECULAR BIOLOGY OBITUARY When the program was finished they estimated that each cycle of provided Carl with the wonderful opportunity to explore ADH struc- refinement would require 16 hours of computing time on the 7090 at ture and function from many perspectives. the IBM computing center at a considerable cost. This type of con- Carl always liked to face grand challenges. In the late 1970s he took strained refinement was well ahead of its time, but methods very sim- on an ambitious project to solve the crystal structure of Rubisco, an ilar to what Brändén and Holmes developed in the early 1960s were enzyme that catalyzes the initial carbon dioxide fixation in bacteria rediscovered by others in the late 1970s and are now used routinely and plants, after reading a report from the US National Academy of for macromolecular crystallographic refinement. Sciences. He first recruited Inger Andersson, a biochemist from In 1962, during Carl’s time at LMB, Crick, Watson and Wilkins won Michael Zeppezauer’s group in Saarbrucken, and later a graduate stu- the Nobel Prize in medicine and Perutz and Kendrew won the Nobel dent, Ylva Lindqvist, whose father was Carl’s Ph.D. thesis mentor. He Prize in chemistry. A celebration party was arranged at Crick’s home, also recruited Gunter Schneider from Hans Eklund’s lab as a postdoc- The Golden Helix, quite memorable for the fireworks arranged by toral fellow. He collaborated with George Lorimer then at DuPont Mark Bretscher. During the party, Perutz made a speech. He said that (Wilmington, Delaware, USA), an expert on the biochemistry of the reason why he and Kendrew had succeeded where others had failed Rubisco, and obtained grams of purified Rubisco protein. Within a few was that they had been so ignorant about crystallography that they did months he had obtained crystals and in 1986 they published the X-ray not know that it was impossible to solve a protein structure. As Carl structure of this bacterial enzyme. This structure became a turning remarked later, “knowledge is fine but too much is inhibitory for point in the studies of Rubisco. Not only did it provide the first breakthrough science.” detailed picture of the active site of Rubisco, but it also provided the After returning to Uppsala in 1963, Carl formed a long-time collab- fold of the large subunit of all Rubisco molecules, which subsequently oration with Hugo Theorell and Åke Åkesson of the Karolinska aided structural and functional interpretation of all Rubisco enzymes. Institute, who were biochemistry experts on alcohol dehydrogenase In collaboration with Stefan Knight, Staffan Normark and Scott (ADH). With an equipment grant Hultgren, Carl determined the from the US National Institutes of crystal structure of a bacterial Health, Brändén took up the struc- chaperone protein, PapD, http://www.nature.com/natstructmolbiol tural determination of this enzyme.